https://sourceforge.net/p/cilib/bugs/2009-04-30T06:28:26ZRecent changes to bugs2009-04-30T06:28:26Z2009-04-30T06:28:26Zhttps://sourceforge.nethttps://sourceforge.neta6b7c3ead1539e951f2f0cfb579c769c9aeec7ea

<div class="markdown_content"><p>The LBestTopology.iterator() method is not implemented, meaning the functionality is inherited from GBestTopology which returns a GBestNeighbourhoodIterator instead of an LBestNeighbourhoodIterator.</p></div>

2009-04-21T07:24:06Z2009-04-21T07:24:06Zhttps://sourceforge.nethttps://sourceforge.net512de1823a0a319082d2739bdc38c377dcb4161f

<div class="markdown_content"><p>The ec.EC algorithm configured with GeneticAlgorithmIterationStrategy is maximising problem that should be minimised.</p></div>

2009-02-13T06:44:40Z2009-02-13T06:44:40Zhttps://sourceforge.nethttps://sourceforge.net2bd069251131c22e989004b8c249fd19f604d3e1

<div class="markdown_content"><p>Someone commented out the FitnessCalculator in the clustering.KMeans class, because FitnessCalculator now requires an Entity and KMeans only has a Vector. I don't recall when this change in FitnessCalculator was made. The Git logs don't contain the change.</p> <p>KMeans is not a population based algorithm and therefore an Entity would not make sense. Is there another type of FitnessCalculator that can handle Vectors?</p> <p>Because the fitness is not calculated, the clusters never get created and the following exception occurs:</p> <p>./simulator.sh xml/kmeans.xml<br /> Starting simulation 1 of 1.<br /> log4j:WARN No appenders could be found for logger (net.sourceforge.cilib.util.ClusteringUtils).<br /> log4j:WARN Please initialize the log4j system properly.<br /> clusters: null<br /> Exception in thread "Thread-1" java.lang.NullPointerException<br /> at net.sourceforge.cilib.clustering.kmeans.KMeans.algorithmIteration(KMeans.java:113)<br /> at net.sourceforge.cilib.algorithm.Algorithm.performIteration(Algorithm.java:141)<br /> at net.sourceforge.cilib.algorithm.Algorithm.run(Algorithm.java:183)<br /> at java.lang.Thread.run(Thread.java:619)</p></div>

2008-10-08T23:11:52Z2008-10-08T23:11:52ZOlusegunhttps://sourceforge.net/u/killearnith/https://sourceforge.net9b96b792f0d86842db06d41681048649d97c044d

<div class="markdown_content"><p>The class doesn't work... Likely because the list of random numbers is not being correctly instantiated.</p> <p>A file that works is attached...</p></div>

2008-10-03T16:16:17Z2008-10-03T16:16:17ZGary Pamparahttps://sourceforge.net/u/gpampara/https://sourceforge.netbe3055683e5d304984a2d74b48278a1931e45c4e

<div class="markdown_content"><p>When the scenario of all entities converging onto a single point in the search space, the .equals() and .hashCode() methods for Entity objects fail. All objects are suddenly the same. This causes a problem within the parentSelection of the DE, resulting in an infinite loop.</p></div>

2008-09-12T14:47:55Z2008-09-12T14:47:55ZOlusegunhttps://sourceforge.net/u/killearnith/https://sourceforge.net02361a14b2dea8dd8dfff41e5ede1f2866fb9a28

<div class="markdown_content"><p>There are two different implementations of the boundary constraint handling mechanisms for pso.</p> <p>One of them defines an enforce(...) interface and can be found in the pso.iterationstrategies package while the other defines a constrainLower(...) and constrainUpper(...) interafce and can be found in pso.positionupdatestrategies.boundaryconstraintstrategies...</p> <p>The former is built into the iteration strategy and provides a more "elegant" solution (this is my opinion). I, thus, suggest that the implementation of the latter be merged into it...</p> <p>Regards,</p> <p>Arnith</p></div>

2008-08-29T15:59:53Z2008-08-29T15:59:53ZOlusegunhttps://sourceforge.net/u/killearnith/https://sourceforge.net46d1c58aa56c777bae2c9de91efed4541079840d

<div class="markdown_content"><p>The strategy should return the winner of the tournament, right? As far as I can tell, this is the individual with the best fitness...</p> <p>The comparator in the select method sorts according to descending fitness (i.e. from worst to best) and returns the worst (individual at index 0)...</p> <p>Suggested fix: comparator should be changed from DescendingFitness to AscendingFitness...</p> <p>I doubt I'm missing anything here but pardon me if I am...</p> <p>Regards...</p> <p>___</p> <p>Killearnith</p></div>

2008-08-28T19:24:34Z2008-08-28T19:24:34ZOlusegunhttps://sourceforge.net/u/killearnith/https://sourceforge.netf332eab572d8f730f42a738eb86f8f1a8b2e9ba8

<div class="markdown_content"><p>The velocity update in question either causes stagnation on Ackley or very slow minimization.</p> <p>Original paper proposing the strategy should be pulled up and the velocity strategy should be re-implemented...</p> <p>__</p> <p>Killearnith</p></div>

2008-08-19T09:17:26Z2008-08-19T09:17:26ZOlusegunhttps://sourceforge.net/u/killearnith/https://sourceforge.net8e5d13adbe810640a06e922746489447059a8b21

<div class="markdown_content"><p>This class is supposed to update part of a cooperative entity (context vector) with the solution represented by participant algorithm entity only if the updated context is better than the original.</p> <p>Currently, the comparison is between the participant entity and the context instead of between the original context and the updated context. This needs to be corrected and tested.</p></div>

2008-06-22T21:12:21Z2008-06-22T21:12:21ZGary Pamparahttps://sourceforge.net/u/gpampara/https://sourceforge.netb4dd44c71e69bdb1041e03e0ec6dc2cf86ac5ce2

<div class="markdown_content"><p>Within a number of objects, the equals() and hashCode() methods are not correct.</p> <p>Many of the equals() methods do not test for reference equality or nulls and the hashCode() should be written to match the equals().</p></div>